China Patent Examiner Cites DeepSeek: AI Assisting Human Judgment in Official Actions?

An examiner in China recently issued a Office Action where, after initially deeming a mathematical formula unclear due to insufficient explanation of its technical rationale, additionally cited an AI tool, DeepSeek, to question the formula’s reasonableness.

Image from the original Medium article.
Image from the original Medium article.

This move, where DeepSeek’s analysis apparently found the formula “unreasonable,” has ignited considerable discussion.

Notably, the Office Action was reportedly taken offline shortly after its issuance.

This incident brings several critical questions for the IP community to the forefront: *The Role of AI: Is this AI assisting examination, or is it an overreach into the examiner’s substantive decision-making?

*Equivalency to “A Person Skilled in the Art”: Can current AI outputs truly replicate the nuanced understanding and judgment expected from a human expert in the relevant technical field?

Is there a responsible pathway for leveraging current AI in patent examination, or should its role be strictly limited to background support for now?

Source: https://lnkd.in/gUr3We2U